Figure 1

Li $1s$ core-level photoemission spectra as a function of temperature. Upper Panel: bulk solid phase. Lower Panel: bulk liquid phase. Vertical dotted lines are guides for the eye. Because of residual water in the experimental chamber, an oxide peak appears in the low temperature spectra as a consequence of a partial surface oxidation. The presence of oxide (accounted for using a Gaussian peak in the fitting procedure) does not affect the results of this report.Reuse & Permissions

Figure 2

Upper panel: Temperature dependence of the squared Gaussian contribution (phonon broadening) to the core-level linewidth for bulk (circles) and surface (squares) peaks. The corresponding best fits of core-level phonon broadening theory—Eq. (1) and Eq. (5) for bulk and surface, respectively—are superimposed as solid lines. Lower Panel: The thermal shift of the bulk peak binding energy.Reuse & Permissions

Figure 3

Upper Panel: (a) Bulk internal energy as measured via calorimetry (solid line) [1], as obtained from phonon broadening (dots) and binding energy shift (open triangles) of the bulk component in our temperature-dependent photoemission data. Lower Panel: (b) Surface internal energy obtained from the phonon broadening of the surface peak (squares) using Eq. (6). The solid line is the theoretical surface internal energy, calculated with Eq. (4). For comparison the bulk internal energy as measured via calorimetry is also shown (thin gray line). The internal energy value at $T=0$ has been subtracted for the sake of comparison.Reuse & Permissions